Genome-wide association study identifies gastric cancer susceptibility loci at 12q24.11-12 and 20q11.21

Abstract

Gastric cancer is the third leading cause of cancer mortality in Japan and worldwide. Although previous studies identify various genetic variations associated with gastric cancer, host genetic factors are largely unidentified. To identify novel gastric cancer loci in the Japanese population, herein, we carried out a large-scale genome-wide association study using 6171 cases and 27 178 controls followed by three replication analyses. Analysis using a total of 11 507 cases and 38 904 controls identified two novel loci on 12q24.11-12 (rs6490061, P = 3.20 × 10^-8 with an odds ratio [OR] of 0.905) and 20q11.21 (rs2376549, P = 8.11 × 10^-10 with an OR of 1.109). rs6490061 is located at intron 19 of the CUX2 gene, and its expression was suppressed by Helicobacter pylori infection. rs2376549 is included within the gene cluster of DEFB families that encode antibacterial peptides. We also found a significant association of rs7849280 in the ABO gene locus on 9q34.2 (P = 2.64 × 10^-13 with an OR of 1.148). CUX2 and ABO expression in gastric mucosal tissues was significantly associated with rs6490061 and rs7849280 (P = 0.0153 and 8.00 × 10^-11 ), respectively. Our findings show the crucial roles of genetic variations in the pathogenesis of gastric cancer.

Keywords: ABO; DEFB; CUX2; gastric cancer; genome-wide association study.

Figure 1

Manhattan plot showing genome‐wide P values of association. The genome‐wide P values of 6 573 681 autosomal single nucleotide polymorphism (SNP) in 6171 cases and 27 178 controls from the screening phase are shown. Three previously reported loci (1q31.1, 5p13.1, and 8q24.3) and five novel loci indicate a significant association. Red horizontal line represents the genome‐wide significance threshold of P = 5.0 × 10⁻⁸. Association of the SNP with gastric cancer risk was investigated by logistic regression analysis using PC1 and PC2 as covariates

Figure 2

Regional plots of three loci for gastric cancer. The −log₁₀ P values from the screening stage in (A) 9q34.2 (rs7849280), (B) 12q24.11‐12 (rs6490061), and (C) 20q11.21 (rs2376549) are shown. Single nucleotide polymorphisms (SNP) genotyped in the replication stage are shown in the figure. Estimated recombination rates (from 1000 Genomes) are plotted in blue. The SNP are color coded to reflect their correlation with the genotyped SNP. Pairwise r ² values are from 1000 Genomes East Asian data (March 2012 release). The genes, position of the exons and direction of transcription are noted. The plots were generated using LocusZoom (http://csg.sph.umich.edu/locuszoom)

Figure 3

Regulation of CUX2 expression by Helicobacter pylori infection and host genetic factors. Box plots indicate the qRT‐PCR analysis of CUX2 mRNA levels in the biopsy samples from the background gastric mucosa. Vertical axis indicates the expression level of CUX2 normalized against GAPDH expression. Box, 25th and 75th percentiles; middle line in the box, median; whiskers, min value inside the 25th percentile − 1.5 × interquartile range and max value inside the 75th percentile + 1.5 × interquartile range; points, outliers. A, CUX2 mRNA in H. pylori‐negative controls (n = 28) and H. pylori‐infected patients (n = 280). B, CUX2 expression levels in each individual patient (n = 53) before and after H. pylori eradication. C, Association of rs6490061 with CUX2 expression in H. pylori‐infected patients (CC, n = 143; TC, n = 117; TT, n = 20). P values were calculated by a t test (A, B) or Kruskal‐Wallis test (C). CC, TC, and TT are genotpe at rs6490061